Spectrum of equivariant cohomology as a fixed point scheme

Basic Information

• Paper ID: 2212.11836
• Title: Spectrum of equivariant cohomology as a fixed point scheme
• Authors: Tamás Hausel, Kamil Rychlewicz (Institute of Science and Technology Austria)
• Classification: math.AG (Algebraic Geometry), math.AT (Algebraic Topology)
• Published Journal: Épijournal de Géométrie Algébrique, Volume 9 (2025), Article No. 1
• Paper Link: https://arxiv.org/abs/2212.11836

Abstract

This paper investigates the geometric realization of the equivariant cohomology ring of a smooth projective variety X acted upon by a complex reductive group G. When the G-action is regular (i.e., all regular unipotent elements have finitely many fixed points), the complex G-equivariant cohomology ring is isomorphic to the coordinate ring of a certain regular fixed point scheme. Principal examples include partial flag varieties, smooth Schubert varieties, and Bott-Samelson varieties. The paper also proves that more general versions of fixed point schemes allow generalization to GKM spaces, such as toric varieties.

Research Background and Motivation

1. Core Problem: How can one geometrically understand the structure of the equivariant cohomology ring? Traditionally, equivariant cohomology is defined through algebraic methods, lacking intuitive geometric interpretation.
2. Problem Significance: Equivariant cohomology has important applications in algebraic geometry, representation theory, and mathematical physics, particularly in the study of Hitchin systems and mirror symmetry. Establishing direct connections between equivariant cohomology and geometric objects facilitates deeper understanding of these theories.
3. Limitations of Existing Methods:
   • The Carrell-Liebermann theorem provides partial results but applies only to specific group actions
   • Lack of a unified geometric framework for understanding equivariant cohomology of general reductive groups
4. Research Motivation:
   • The authors discovered while studying the Hitchin map on Grassmannians that a certain infinitesimal fixed point scheme is isomorphic to the spectrum of equivariant cohomology
   • This inspired the search for geometric realizations of equivariant cohomology in more general settings

Core Contributions

1. Main Theorem: Proves that for a regular action of a principally paired group H on a smooth projective variety X, the coordinate ring of the zero scheme $Z_S \subset S \times X$ (where S is a Kostant section) is isomorphic to the H-equivariant cohomology ring.
2. Unified Framework: Establishes a unified theoretical framework extending from solvable groups to reductive groups to general principally paired groups.
3. Geometric Realization: Concretely realizes the abstract equivariant cohomology ring as the coordinate ring of an affine scheme, providing intuitive geometric interpretation.
4. Generalization Results:
   • Proves similar results hold for GKM spaces (such as toric varieties)
   • Establishes the relationship between the total zero scheme and equivariant cohomology
5. Functoriality: Proves the functoriality of the construction, i.e., for equivariant maps, the corresponding scheme maps are compatible with pullback maps in equivariant cohomology.

Detailed Methodology

Problem Setup

Given a principally paired group H acting on a smooth projective variety X, construct an affine scheme whose coordinate ring is isomorphic to the H-equivariant cohomology ring $H_H^*(X;\mathbb{C})$.

Core Concepts

1. Principally Paired Groups

Definition: A complex linear algebraic group H is called principally paired if its Lie algebra $\mathfrak{h}$ contains a pair $(e,h) \subset \mathfrak{h}$ satisfying:

• $[h,e] = 2e$ and e is a regular unipotent element
• There exists an algebraic group homomorphism $B(SL_2) \to H$ whose differential maps the regular unipotent element to e

Examples:

• Reductive groups (via the Jacobson-Morozov theorem)
• Parabolic subgroups

2. Regular Actions

Definition: An action of a principally paired group H on a smooth projective variety X is called regular if regular unipotent elements $u \in H$ have finitely many fixed points.

Since unipotent elements always have connected fixed point sets, regular action means $X^u = \{o\}$ for some $o \in X$.

3. Generalization of Kostant Sections

For a general principally paired group H, choose a Levi subgroup $L \subset H$ such that $H = N \rtimes L$, where N is the unipotent radical of H. The regular unipotent element $e \in H$ decomposes as $e = e_n + e_l$, where $e_n \in \mathfrak{n}, e_l \in \mathfrak{l}$.

The generalized Kostant section is defined as: $S := e + C_{\mathfrak{l}}(f_l) \subset \mathfrak{h}$

where $(e_l, f_l, h_l)$ is an $sl_2$-triple in $\mathfrak{l}$.

Model Architecture

1. Total Vector Field Construction

Define a total vector field $V_{\mathfrak{h}}$ on $\mathfrak{h} \times X$ such that for any $y \in \mathfrak{h}$, its restriction $(V_{\mathfrak{h}})_y$ to $\{y\} \times X$ is the infinitesimal vector field generated by y.

2. Zero Scheme Definition

Let $V_S := V_{\mathfrak{h}}|_{S \times X}$ be the restriction of the total vector field to $S \times X$. The zero scheme $Z_S \subset S \times X$ is defined as the zero scheme of $V_S$, i.e., the subscheme defined by the ideal sheaf $V_S(\mathcal{O}_{S \times X}) \subset \mathcal{O}_{S \times X}$.

3. $\mathbb{C}^*$-Action

Through the principal $sl_2$-subgroup $\mathbb{C}^* \to H$, define a $\mathbb{C}^*$-action on $S \times X$:

• On S: $t \cdot v = t^{-2}\text{Ad}_{H^t}(v)$
• On X: through the action of $H^t$

Technical Innovations

1. Unified Approach

By introducing the concept of principally paired groups, the paper unifies treatment of various cases from solvable groups to reductive groups, avoiding case-by-case analysis.

2. Vector Field Technique

Cleverly uses the zero schemes of vector fields to encode equivariant cohomology information, providing more intuitive understanding than traditional algebraic methods.

3. Graded Structure

Naturally obtains a graded structure through the $\mathbb{C}^*$-action, ensuring that the algebraic isomorphism preserves grading.

Main Theorems

Theorem 1.2 (Main Theorem)

Let a principally paired group H act regularly on a smooth projective complex variety X. Then the zero scheme $Z_S \subset S \times X$ is a reduced affine scheme whose coordinate ring as a graded ring is isomorphic to H-equivariant cohomology:

$\mathbb{C}[Z_S] \cong H_H^*(X;\mathbb{C}) \text{ over } \mathbb{C}[S] \cong H_H^*$

In particular, there is an isomorphism: $Z_S \cong \text{Spec}(H_H^*(X;\mathbb{C})), \quad S \cong \text{Spec}(H_H^*)$

Theorem 1.3 (Total Zero Scheme Version)

For a reductive group G acting regularly on X, the ring of G-invariant functions on the total zero scheme $Z_{\mathfrak{g}} \subset \mathfrak{g} \times X$ is isomorphic to equivariant cohomology:

$\mathbb{C}[Z_{\mathfrak{g}}]^G \cong H_G^*(X;\mathbb{C}) \text{ over } \mathbb{C}[\mathfrak{g}]^G \cong H_G^*$

Theorem 1.4 (GKM Spaces)

Let a torus T act on a smooth projective complex variety X such that X is a GKM space (i.e., has finitely many 0-dimensional and 1-dimensional orbits). Then:

$\mathbb{C}[Z_{\mathfrak{t}}] \cong H_T^*(X;\mathbb{C}) \text{ over } \mathbb{C}[\mathfrak{t}] \cong H_T^*$

Proof Strategy

1. Solvable Group Case (Section 3)

• Utilize Białynicki-Birula decomposition and equivariant formality
• Prove the zero scheme is a complete intersection and reduced
• Establish isomorphism through localization

2. Reductive Group Case (Section 4)

• Reduce to the Borel subgroup case
• Utilize the quotient structure of Weyl group action
• Use properties of Kostant sections

3. General Principally Paired Groups (Section 4.5)

• Combine techniques from solvable and reductive groups
• Utilize Levi decomposition

Important Examples

1. Projective Space

For $SL_2$ acting on $\mathbb{P}^n$: $H^*_{SL_2}(\mathbb{P}^n) \cong \mathbb{C}[t,x]/I$ where I is generated by the polynomial $x(x+2t)(x+4t)\cdots(x+2nt)$.

2. Grassmannian

For $SL_2$ acting on $Gr(2,4)$, the equivariant cohomology ring can be explicitly computed, with the corresponding scheme having six one-parameter families of solutions.

3. Flag Varieties

For the complete flag variety $F_3 = SL_3/B$, the geometric realization of the equivariant cohomology ring provides rich combinatorial structure.

4. Bott-Samelson Varieties

As resolutions of Schubert varieties, Bott-Samelson varieties provide important examples of regular B-varieties.

Applications and Generalizations

1. Hitchin Systems

The geometric realization of equivariant cohomology rings provides new perspectives for understanding the behavior of Hitchin systems on various Lagrangian submanifolds.

2. Generalization to Singular Varieties

Through appropriate conditions (such as surjectivity of cohomology), results can be generalized to certain singular varieties.

3. GKM Theory

For GKM spaces with torus actions, provides new geometric interpretation of equivariant cohomology.

Technical Tools

1. Vector Field Theory

• Generalization of the Carrell-Liebermann theorem
• Analysis of vector field zeros

2. Algebraic Group Theory

• Jordan decomposition
• Levi decomposition
• Properties of Borel subgroups

3. Equivariant Cohomology

• Localization theorem
• Equivariant formality
• Equivariant versions of Chern classes

Conclusions and Perspectives

Main Contributions

1. Establishes direct connections between equivariant cohomology rings and geometric objects (zero schemes)
2. Provides a unified theoretical framework for handling various group actions
3. Furnishes new geometric tools for applications such as Hitchin systems

Theoretical Significance

• Deepens understanding of the geometric nature of equivariant cohomology
• Provides new methods for cross-disciplinary research between algebraic geometry and representation theory
• Connects vector field theory with equivariant topology

Future Directions

1. Generalize to more general group actions and spaces
2. Explore connections with mirror symmetry
3. Develop analogous theory for equivariant K-theory
4. Applications in mathematical physics

Through clever geometric construction, this paper provides new perspectives on equivariant cohomology theory, with significant implications in the intersection of algebraic geometry, representation theory, and mathematical physics.